Prevalence assessment adjusted for laboratory test performance using an example of the COVID-19 serological tests

Abstract

Assessment of the prevalence of the disease or condition should consider the accuracy of the diagnostic tests. In the context of the new coronavirus infection (COVID-19) pandemic, laboratory testing has been one of the most important components of the overall strategy for the control and prevention of this infection. Seroprevalence studies have been used to assess and monitor the level of population immunity to the virus.

In this paper we provide detailed description of the methods to calculate and interpret the accuracy of laboratory tests as well as their sensitivity, specificity, positive- and negative prognostic values of laboratory tests using seroprevalence of COVID-19 studies as an example for better understanding of the methodological issues. The use of the laboratory tests accuracy in prevalence studies has been demonstrated. A sample syntax to calculate confidence intervals for the prevalence estimates using the bootstrap procedure with known absolute values of true positive and true negative results, false positive and false negative results for R software is also provided. Presentation of the prevalence estimates adjusted for test performance indicators with confidence intervals improves comparability of the findings obtained using different serological tests.

The article is intended for undergraduate-, postgraduate-, and doctoral students in health sciences working with the assessment of the prevalence (seroprevalence) of diseases or conditions through population-based serological surveys.